1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
|
;
; Copyright (c) 2010 The WebM project authors. All Rights Reserved.
;
; Use of this source code is governed by a BSD-style license
; that can be found in the LICENSE file in the root of the source
; tree. An additional intellectual property rights grant can be found
; in the file PATENTS. All contributing project authors may
; be found in the AUTHORS file in the root of the source tree.
;
EXPORT |vp8_loop_filter_horizontal_edge_armv6|
EXPORT |vp8_mbloop_filter_horizontal_edge_armv6|
EXPORT |vp8_loop_filter_vertical_edge_armv6|
EXPORT |vp8_mbloop_filter_vertical_edge_armv6|
AREA |.text|, CODE, READONLY ; name this block of code
MACRO
TRANSPOSE_MATRIX $a0, $a1, $a2, $a3, $b0, $b1, $b2, $b3
; input: $a0, $a1, $a2, $a3; output: $b0, $b1, $b2, $b3
; a0: 03 02 01 00
; a1: 13 12 11 10
; a2: 23 22 21 20
; a3: 33 32 31 30
; b3 b2 b1 b0
uxtb16 $b1, $a1 ; xx 12 xx 10
uxtb16 $b0, $a0 ; xx 02 xx 00
uxtb16 $b3, $a3 ; xx 32 xx 30
uxtb16 $b2, $a2 ; xx 22 xx 20
orr $b1, $b0, $b1, lsl #8 ; 12 02 10 00
orr $b3, $b2, $b3, lsl #8 ; 32 22 30 20
uxtb16 $a1, $a1, ror #8 ; xx 13 xx 11
uxtb16 $a3, $a3, ror #8 ; xx 33 xx 31
uxtb16 $a0, $a0, ror #8 ; xx 03 xx 01
uxtb16 $a2, $a2, ror #8 ; xx 23 xx 21
orr $a0, $a0, $a1, lsl #8 ; 13 03 11 01
orr $a2, $a2, $a3, lsl #8 ; 33 23 31 21
pkhtb $b2, $b3, $b1, asr #16 ; 32 22 12 02 -- p1
pkhbt $b0, $b1, $b3, lsl #16 ; 30 20 10 00 -- p3
pkhtb $b3, $a2, $a0, asr #16 ; 33 23 13 03 -- p0
pkhbt $b1, $a0, $a2, lsl #16 ; 31 21 11 01 -- p2
MEND
src RN r0
pstep RN r1
count RN r5
;r0 unsigned char *src_ptr,
;r1 int src_pixel_step,
;r2 const char *blimit,
;r3 const char *limit,
;stack const char *thresh,
;stack int count
;-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|vp8_loop_filter_horizontal_edge_armv6| PROC
;-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
stmdb sp!, {r4 - r11, lr}
sub src, src, pstep, lsl #2 ; move src pointer down by 4 lines
ldr count, [sp, #40] ; count for 8-in-parallel
ldr r6, [sp, #36] ; load thresh address
sub sp, sp, #16 ; create temp buffer
ldr r9, [src], pstep ; p3
ldrb r4, [r2] ; blimit
ldr r10, [src], pstep ; p2
ldrb r2, [r3] ; limit
ldr r11, [src], pstep ; p1
orr r4, r4, r4, lsl #8
ldrb r3, [r6] ; thresh
orr r2, r2, r2, lsl #8
mov count, count, lsl #1 ; 4-in-parallel
orr r4, r4, r4, lsl #16
orr r3, r3, r3, lsl #8
orr r2, r2, r2, lsl #16
orr r3, r3, r3, lsl #16
|Hnext8|
; vp8_filter_mask() function
; calculate breakout conditions
ldr r12, [src], pstep ; p0
uqsub8 r6, r9, r10 ; p3 - p2
uqsub8 r7, r10, r9 ; p2 - p3
uqsub8 r8, r10, r11 ; p2 - p1
uqsub8 r10, r11, r10 ; p1 - p2
orr r6, r6, r7 ; abs (p3-p2)
orr r8, r8, r10 ; abs (p2-p1)
uqsub8 lr, r6, r2 ; compare to limit. lr: vp8_filter_mask
uqsub8 r8, r8, r2 ; compare to limit
uqsub8 r6, r11, r12 ; p1 - p0
orr lr, lr, r8
uqsub8 r7, r12, r11 ; p0 - p1
ldr r9, [src], pstep ; q0
ldr r10, [src], pstep ; q1
orr r6, r6, r7 ; abs (p1-p0)
uqsub8 r7, r6, r2 ; compare to limit
uqsub8 r8, r6, r3 ; compare to thresh -- save r8 for later
orr lr, lr, r7
uqsub8 r6, r11, r10 ; p1 - q1
uqsub8 r7, r10, r11 ; q1 - p1
uqsub8 r11, r12, r9 ; p0 - q0
uqsub8 r12, r9, r12 ; q0 - p0
orr r6, r6, r7 ; abs (p1-q1)
ldr r7, c0x7F7F7F7F
orr r12, r11, r12 ; abs (p0-q0)
ldr r11, [src], pstep ; q2
uqadd8 r12, r12, r12 ; abs (p0-q0) * 2
and r6, r7, r6, lsr #1 ; abs (p1-q1) / 2
uqsub8 r7, r9, r10 ; q0 - q1
uqadd8 r12, r12, r6 ; abs (p0-q0)*2 + abs (p1-q1)/2
uqsub8 r6, r10, r9 ; q1 - q0
uqsub8 r12, r12, r4 ; compare to flimit
uqsub8 r9, r11, r10 ; q2 - q1
orr lr, lr, r12
ldr r12, [src], pstep ; q3
uqsub8 r10, r10, r11 ; q1 - q2
orr r6, r7, r6 ; abs (q1-q0)
orr r10, r9, r10 ; abs (q2-q1)
uqsub8 r7, r6, r2 ; compare to limit
uqsub8 r10, r10, r2 ; compare to limit
uqsub8 r6, r6, r3 ; compare to thresh -- save r6 for later
orr lr, lr, r7
orr lr, lr, r10
uqsub8 r10, r12, r11 ; q3 - q2
uqsub8 r9, r11, r12 ; q2 - q3
mvn r11, #0 ; r11 == -1
orr r10, r10, r9 ; abs (q3-q2)
uqsub8 r10, r10, r2 ; compare to limit
mov r12, #0
orr lr, lr, r10
sub src, src, pstep, lsl #2
usub8 lr, r12, lr ; use usub8 instead of ssub8
sel lr, r11, r12 ; filter mask: lr
cmp lr, #0
beq hskip_filter ; skip filtering
sub src, src, pstep, lsl #1 ; move src pointer down by 6 lines
;vp8_hevmask() function
;calculate high edge variance
orr r10, r6, r8 ; calculate vp8_hevmask
ldr r7, [src], pstep ; p1
usub8 r10, r12, r10 ; use usub8 instead of ssub8
sel r6, r12, r11 ; obtain vp8_hevmask: r6
;vp8_filter() function
ldr r8, [src], pstep ; p0
ldr r12, c0x80808080
ldr r9, [src], pstep ; q0
ldr r10, [src], pstep ; q1
eor r7, r7, r12 ; p1 offset to convert to a signed value
eor r8, r8, r12 ; p0 offset to convert to a signed value
eor r9, r9, r12 ; q0 offset to convert to a signed value
eor r10, r10, r12 ; q1 offset to convert to a signed value
str r9, [sp] ; store qs0 temporarily
str r8, [sp, #4] ; store ps0 temporarily
str r10, [sp, #8] ; store qs1 temporarily
str r7, [sp, #12] ; store ps1 temporarily
qsub8 r7, r7, r10 ; vp8_signed_char_clamp(ps1-qs1)
qsub8 r8, r9, r8 ; vp8_signed_char_clamp(vp8_filter + 3 * ( qs0 - ps0))
and r7, r7, r6 ; vp8_filter (r7) &= hev
qadd8 r7, r7, r8
ldr r9, c0x03030303 ; r9 = 3 --modified for vp8
qadd8 r7, r7, r8
ldr r10, c0x04040404
qadd8 r7, r7, r8
and r7, r7, lr ; vp8_filter &= mask;
;modify code for vp8 -- Filter1 = vp8_filter (r7)
qadd8 r8 , r7 , r9 ; Filter2 (r8) = vp8_signed_char_clamp(vp8_filter+3)
qadd8 r7 , r7 , r10 ; vp8_filter = vp8_signed_char_clamp(vp8_filter+4)
mov r9, #0
shadd8 r8 , r8 , r9 ; Filter2 >>= 3
shadd8 r7 , r7 , r9 ; vp8_filter >>= 3
shadd8 r8 , r8 , r9
shadd8 r7 , r7 , r9
shadd8 lr , r8 , r9 ; lr: Filter2
shadd8 r7 , r7 , r9 ; r7: filter
;usub8 lr, r8, r10 ; s = (s==4)*-1
;sel lr, r11, r9
;usub8 r8, r10, r8
;sel r8, r11, r9
;and r8, r8, lr ; -1 for each element that equals 4
;calculate output
;qadd8 lr, r8, r7 ; u = vp8_signed_char_clamp(s + vp8_filter)
ldr r8, [sp] ; load qs0
ldr r9, [sp, #4] ; load ps0
ldr r10, c0x01010101
qsub8 r8 ,r8, r7 ; u = vp8_signed_char_clamp(qs0 - vp8_filter)
qadd8 r9, r9, lr ; u = vp8_signed_char_clamp(ps0 + Filter2)
;end of modification for vp8
mov lr, #0
sadd8 r7, r7 , r10 ; vp8_filter += 1
shadd8 r7, r7, lr ; vp8_filter >>= 1
ldr r11, [sp, #12] ; load ps1
ldr r10, [sp, #8] ; load qs1
bic r7, r7, r6 ; vp8_filter &= ~hev
sub src, src, pstep, lsl #2
qadd8 r11, r11, r7 ; u = vp8_signed_char_clamp(ps1 + vp8_filter)
qsub8 r10, r10,r7 ; u = vp8_signed_char_clamp(qs1 - vp8_filter)
eor r11, r11, r12 ; *op1 = u^0x80
str r11, [src], pstep ; store op1
eor r9, r9, r12 ; *op0 = u^0x80
str r9, [src], pstep ; store op0 result
eor r8, r8, r12 ; *oq0 = u^0x80
str r8, [src], pstep ; store oq0 result
eor r10, r10, r12 ; *oq1 = u^0x80
str r10, [src], pstep ; store oq1
sub src, src, pstep, lsl #1
|hskip_filter|
add src, src, #4
sub src, src, pstep, lsl #2
subs count, count, #1
ldrne r9, [src], pstep ; p3
ldrne r10, [src], pstep ; p2
ldrne r11, [src], pstep ; p1
bne Hnext8
add sp, sp, #16
ldmia sp!, {r4 - r11, pc}
ENDP ; |vp8_loop_filter_horizontal_edge_armv6|
;-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|vp8_mbloop_filter_horizontal_edge_armv6| PROC
;-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
stmdb sp!, {r4 - r11, lr}
sub src, src, pstep, lsl #2 ; move src pointer down by 4 lines
ldr count, [sp, #40] ; count for 8-in-parallel
ldr r6, [sp, #36] ; load thresh address
sub sp, sp, #16 ; create temp buffer
ldr r9, [src], pstep ; p3
ldrb r4, [r2] ; blimit
ldr r10, [src], pstep ; p2
ldrb r2, [r3] ; limit
ldr r11, [src], pstep ; p1
orr r4, r4, r4, lsl #8
ldrb r3, [r6] ; thresh
orr r2, r2, r2, lsl #8
mov count, count, lsl #1 ; 4-in-parallel
orr r4, r4, r4, lsl #16
orr r3, r3, r3, lsl #8
orr r2, r2, r2, lsl #16
orr r3, r3, r3, lsl #16
|MBHnext8|
; vp8_filter_mask() function
; calculate breakout conditions
ldr r12, [src], pstep ; p0
uqsub8 r6, r9, r10 ; p3 - p2
uqsub8 r7, r10, r9 ; p2 - p3
uqsub8 r8, r10, r11 ; p2 - p1
uqsub8 r10, r11, r10 ; p1 - p2
orr r6, r6, r7 ; abs (p3-p2)
orr r8, r8, r10 ; abs (p2-p1)
uqsub8 lr, r6, r2 ; compare to limit. lr: vp8_filter_mask
uqsub8 r8, r8, r2 ; compare to limit
uqsub8 r6, r11, r12 ; p1 - p0
orr lr, lr, r8
uqsub8 r7, r12, r11 ; p0 - p1
ldr r9, [src], pstep ; q0
ldr r10, [src], pstep ; q1
orr r6, r6, r7 ; abs (p1-p0)
uqsub8 r7, r6, r2 ; compare to limit
uqsub8 r8, r6, r3 ; compare to thresh -- save r8 for later
orr lr, lr, r7
uqsub8 r6, r11, r10 ; p1 - q1
uqsub8 r7, r10, r11 ; q1 - p1
uqsub8 r11, r12, r9 ; p0 - q0
uqsub8 r12, r9, r12 ; q0 - p0
orr r6, r6, r7 ; abs (p1-q1)
ldr r7, c0x7F7F7F7F
orr r12, r11, r12 ; abs (p0-q0)
ldr r11, [src], pstep ; q2
uqadd8 r12, r12, r12 ; abs (p0-q0) * 2
and r6, r7, r6, lsr #1 ; abs (p1-q1) / 2
uqsub8 r7, r9, r10 ; q0 - q1
uqadd8 r12, r12, r6 ; abs (p0-q0)*2 + abs (p1-q1)/2
uqsub8 r6, r10, r9 ; q1 - q0
uqsub8 r12, r12, r4 ; compare to flimit
uqsub8 r9, r11, r10 ; q2 - q1
orr lr, lr, r12
ldr r12, [src], pstep ; q3
uqsub8 r10, r10, r11 ; q1 - q2
orr r6, r7, r6 ; abs (q1-q0)
orr r10, r9, r10 ; abs (q2-q1)
uqsub8 r7, r6, r2 ; compare to limit
uqsub8 r10, r10, r2 ; compare to limit
uqsub8 r6, r6, r3 ; compare to thresh -- save r6 for later
orr lr, lr, r7
orr lr, lr, r10
uqsub8 r10, r12, r11 ; q3 - q2
uqsub8 r9, r11, r12 ; q2 - q3
mvn r11, #0 ; r11 == -1
orr r10, r10, r9 ; abs (q3-q2)
uqsub8 r10, r10, r2 ; compare to limit
mov r12, #0
orr lr, lr, r10
usub8 lr, r12, lr ; use usub8 instead of ssub8
sel lr, r11, r12 ; filter mask: lr
cmp lr, #0
beq mbhskip_filter ; skip filtering
;vp8_hevmask() function
;calculate high edge variance
sub src, src, pstep, lsl #2 ; move src pointer down by 6 lines
sub src, src, pstep, lsl #1
orr r10, r6, r8
ldr r7, [src], pstep ; p1
usub8 r10, r12, r10
sel r6, r12, r11 ; hev mask: r6
;vp8_mbfilter() function
;p2, q2 are only needed at the end. Don't need to load them in now.
ldr r8, [src], pstep ; p0
ldr r12, c0x80808080
ldr r9, [src], pstep ; q0
ldr r10, [src] ; q1
eor r7, r7, r12 ; ps1
eor r8, r8, r12 ; ps0
eor r9, r9, r12 ; qs0
eor r10, r10, r12 ; qs1
qsub8 r12, r9, r8 ; vp8_signed_char_clamp(vp8_filter + 3 * ( qs0 - ps0))
str r7, [sp, #12] ; store ps1 temporarily
qsub8 r7, r7, r10 ; vp8_signed_char_clamp(ps1-qs1)
str r10, [sp, #8] ; store qs1 temporarily
qadd8 r7, r7, r12
str r9, [sp] ; store qs0 temporarily
qadd8 r7, r7, r12
str r8, [sp, #4] ; store ps0 temporarily
qadd8 r7, r7, r12 ; vp8_filter: r7
ldr r10, c0x03030303 ; r10 = 3 --modified for vp8
ldr r9, c0x04040404
and r7, r7, lr ; vp8_filter &= mask (lr is free)
mov r12, r7 ; Filter2: r12
and r12, r12, r6 ; Filter2 &= hev
;modify code for vp8
;save bottom 3 bits so that we round one side +4 and the other +3
qadd8 r8 , r12 , r9 ; Filter1 (r8) = vp8_signed_char_clamp(Filter2+4)
qadd8 r12 , r12 , r10 ; Filter2 (r12) = vp8_signed_char_clamp(Filter2+3)
mov r10, #0
shadd8 r8 , r8 , r10 ; Filter1 >>= 3
shadd8 r12 , r12 , r10 ; Filter2 >>= 3
shadd8 r8 , r8 , r10
shadd8 r12 , r12 , r10
shadd8 r8 , r8 , r10 ; r8: Filter1
shadd8 r12 , r12 , r10 ; r12: Filter2
ldr r9, [sp] ; load qs0
ldr r11, [sp, #4] ; load ps0
qsub8 r9 , r9, r8 ; qs0 = vp8_signed_char_clamp(qs0 - Filter1)
qadd8 r11, r11, r12 ; ps0 = vp8_signed_char_clamp(ps0 + Filter2)
;save bottom 3 bits so that we round one side +4 and the other +3
;and r8, r12, r10 ; s = Filter2 & 7 (s: r8)
;qadd8 r12 , r12 , r9 ; Filter2 = vp8_signed_char_clamp(Filter2+4)
;mov r10, #0
;shadd8 r12 , r12 , r10 ; Filter2 >>= 3
;usub8 lr, r8, r9 ; s = (s==4)*-1
;sel lr, r11, r10
;shadd8 r12 , r12 , r10
;usub8 r8, r9, r8
;sel r8, r11, r10
;ldr r9, [sp] ; load qs0
;ldr r11, [sp, #4] ; load ps0
;shadd8 r12 , r12 , r10
;and r8, r8, lr ; -1 for each element that equals 4
;qadd8 r10, r8, r12 ; u = vp8_signed_char_clamp(s + Filter2)
;qsub8 r9 , r9, r12 ; qs0 = vp8_signed_char_clamp(qs0 - Filter2)
;qadd8 r11, r11, r10 ; ps0 = vp8_signed_char_clamp(ps0 + u)
;end of modification for vp8
bic r12, r7, r6 ; vp8_filter &= ~hev ( r6 is free)
;mov r12, r7
;roughly 3/7th difference across boundary
mov lr, #0x1b ; 27
mov r7, #0x3f ; 63
sxtb16 r6, r12
sxtb16 r10, r12, ror #8
smlabb r8, r6, lr, r7
smlatb r6, r6, lr, r7
smlabb r7, r10, lr, r7
smultb r10, r10, lr
ssat r8, #8, r8, asr #7
ssat r6, #8, r6, asr #7
add r10, r10, #63
ssat r7, #8, r7, asr #7
ssat r10, #8, r10, asr #7
ldr lr, c0x80808080
pkhbt r6, r8, r6, lsl #16
pkhbt r10, r7, r10, lsl #16
uxtb16 r6, r6
uxtb16 r10, r10
sub src, src, pstep
orr r10, r6, r10, lsl #8 ; u = vp8_signed_char_clamp((63 + Filter2 * 27)>>7)
qsub8 r8, r9, r10 ; s = vp8_signed_char_clamp(qs0 - u)
qadd8 r10, r11, r10 ; s = vp8_signed_char_clamp(ps0 + u)
eor r8, r8, lr ; *oq0 = s^0x80
str r8, [src] ; store *oq0
sub src, src, pstep
eor r10, r10, lr ; *op0 = s^0x80
str r10, [src] ; store *op0
;roughly 2/7th difference across boundary
mov lr, #0x12 ; 18
mov r7, #0x3f ; 63
sxtb16 r6, r12
sxtb16 r10, r12, ror #8
smlabb r8, r6, lr, r7
smlatb r6, r6, lr, r7
smlabb r9, r10, lr, r7
smlatb r10, r10, lr, r7
ssat r8, #8, r8, asr #7
ssat r6, #8, r6, asr #7
ssat r9, #8, r9, asr #7
ssat r10, #8, r10, asr #7
ldr lr, c0x80808080
pkhbt r6, r8, r6, lsl #16
pkhbt r10, r9, r10, lsl #16
ldr r9, [sp, #8] ; load qs1
ldr r11, [sp, #12] ; load ps1
uxtb16 r6, r6
uxtb16 r10, r10
sub src, src, pstep
orr r10, r6, r10, lsl #8 ; u = vp8_signed_char_clamp((63 + Filter2 * 18)>>7)
qadd8 r11, r11, r10 ; s = vp8_signed_char_clamp(ps1 + u)
qsub8 r8, r9, r10 ; s = vp8_signed_char_clamp(qs1 - u)
eor r11, r11, lr ; *op1 = s^0x80
str r11, [src], pstep ; store *op1
eor r8, r8, lr ; *oq1 = s^0x80
add src, src, pstep, lsl #1
mov r7, #0x3f ; 63
str r8, [src], pstep ; store *oq1
;roughly 1/7th difference across boundary
mov lr, #0x9 ; 9
ldr r9, [src] ; load q2
sxtb16 r6, r12
sxtb16 r10, r12, ror #8
smlabb r8, r6, lr, r7
smlatb r6, r6, lr, r7
smlabb r12, r10, lr, r7
smlatb r10, r10, lr, r7
ssat r8, #8, r8, asr #7
ssat r6, #8, r6, asr #7
ssat r12, #8, r12, asr #7
ssat r10, #8, r10, asr #7
sub src, src, pstep, lsl #2
pkhbt r6, r8, r6, lsl #16
pkhbt r10, r12, r10, lsl #16
sub src, src, pstep
ldr lr, c0x80808080
ldr r11, [src] ; load p2
uxtb16 r6, r6
uxtb16 r10, r10
eor r9, r9, lr
eor r11, r11, lr
orr r10, r6, r10, lsl #8 ; u = vp8_signed_char_clamp((63 + Filter2 * 9)>>7)
qadd8 r8, r11, r10 ; s = vp8_signed_char_clamp(ps2 + u)
qsub8 r10, r9, r10 ; s = vp8_signed_char_clamp(qs2 - u)
eor r8, r8, lr ; *op2 = s^0x80
str r8, [src], pstep, lsl #2 ; store *op2
add src, src, pstep
eor r10, r10, lr ; *oq2 = s^0x80
str r10, [src], pstep, lsl #1 ; store *oq2
|mbhskip_filter|
add src, src, #4
sub src, src, pstep, lsl #3
subs count, count, #1
ldrne r9, [src], pstep ; p3
ldrne r10, [src], pstep ; p2
ldrne r11, [src], pstep ; p1
bne MBHnext8
add sp, sp, #16
ldmia sp!, {r4 - r11, pc}
ENDP ; |vp8_mbloop_filter_horizontal_edge_armv6|
;-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|vp8_loop_filter_vertical_edge_armv6| PROC
;-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
stmdb sp!, {r4 - r11, lr}
sub src, src, #4 ; move src pointer down by 4
ldr count, [sp, #40] ; count for 8-in-parallel
ldr r12, [sp, #36] ; load thresh address
sub sp, sp, #16 ; create temp buffer
ldr r6, [src], pstep ; load source data
ldrb r4, [r2] ; blimit
ldr r7, [src], pstep
ldrb r2, [r3] ; limit
ldr r8, [src], pstep
orr r4, r4, r4, lsl #8
ldrb r3, [r12] ; thresh
orr r2, r2, r2, lsl #8
ldr lr, [src], pstep
mov count, count, lsl #1 ; 4-in-parallel
orr r4, r4, r4, lsl #16
orr r3, r3, r3, lsl #8
orr r2, r2, r2, lsl #16
orr r3, r3, r3, lsl #16
|Vnext8|
; vp8_filter_mask() function
; calculate breakout conditions
; transpose the source data for 4-in-parallel operation
TRANSPOSE_MATRIX r6, r7, r8, lr, r9, r10, r11, r12
uqsub8 r7, r9, r10 ; p3 - p2
uqsub8 r8, r10, r9 ; p2 - p3
uqsub8 r9, r10, r11 ; p2 - p1
uqsub8 r10, r11, r10 ; p1 - p2
orr r7, r7, r8 ; abs (p3-p2)
orr r10, r9, r10 ; abs (p2-p1)
uqsub8 lr, r7, r2 ; compare to limit. lr: vp8_filter_mask
uqsub8 r10, r10, r2 ; compare to limit
sub src, src, pstep, lsl #2 ; move src pointer down by 4 lines
orr lr, lr, r10
uqsub8 r6, r11, r12 ; p1 - p0
uqsub8 r7, r12, r11 ; p0 - p1
add src, src, #4 ; move src pointer up by 4
orr r6, r6, r7 ; abs (p1-p0)
str r11, [sp, #12] ; save p1
uqsub8 r10, r6, r2 ; compare to limit
uqsub8 r11, r6, r3 ; compare to thresh
orr lr, lr, r10
; transpose uses 8 regs(r6 - r12 and lr). Need to save reg value now
; transpose the source data for 4-in-parallel operation
ldr r6, [src], pstep ; load source data
str r11, [sp] ; push r11 to stack
ldr r7, [src], pstep
str r12, [sp, #4] ; save current reg before load q0 - q3 data
ldr r8, [src], pstep
str lr, [sp, #8]
ldr lr, [src], pstep
TRANSPOSE_MATRIX r6, r7, r8, lr, r9, r10, r11, r12
ldr lr, [sp, #8] ; load back (f)limit accumulator
uqsub8 r6, r12, r11 ; q3 - q2
uqsub8 r7, r11, r12 ; q2 - q3
uqsub8 r12, r11, r10 ; q2 - q1
uqsub8 r11, r10, r11 ; q1 - q2
orr r6, r6, r7 ; abs (q3-q2)
orr r7, r12, r11 ; abs (q2-q1)
uqsub8 r6, r6, r2 ; compare to limit
uqsub8 r7, r7, r2 ; compare to limit
ldr r11, [sp, #4] ; load back p0
ldr r12, [sp, #12] ; load back p1
orr lr, lr, r6
orr lr, lr, r7
uqsub8 r6, r11, r9 ; p0 - q0
uqsub8 r7, r9, r11 ; q0 - p0
uqsub8 r8, r12, r10 ; p1 - q1
uqsub8 r11, r10, r12 ; q1 - p1
orr r6, r6, r7 ; abs (p0-q0)
ldr r7, c0x7F7F7F7F
orr r8, r8, r11 ; abs (p1-q1)
uqadd8 r6, r6, r6 ; abs (p0-q0) * 2
and r8, r7, r8, lsr #1 ; abs (p1-q1) / 2
uqsub8 r11, r10, r9 ; q1 - q0
uqadd8 r6, r8, r6 ; abs (p0-q0)*2 + abs (p1-q1)/2
uqsub8 r12, r9, r10 ; q0 - q1
uqsub8 r6, r6, r4 ; compare to flimit
orr r9, r11, r12 ; abs (q1-q0)
uqsub8 r8, r9, r2 ; compare to limit
uqsub8 r10, r9, r3 ; compare to thresh
orr lr, lr, r6
orr lr, lr, r8
mvn r11, #0 ; r11 == -1
mov r12, #0
usub8 lr, r12, lr
ldr r9, [sp] ; load the compared result
sel lr, r11, r12 ; filter mask: lr
cmp lr, #0
beq vskip_filter ; skip filtering
;vp8_hevmask() function
;calculate high edge variance
sub src, src, pstep, lsl #2 ; move src pointer down by 4 lines
orr r9, r9, r10
ldrh r7, [src, #-2]
ldrh r8, [src], pstep
usub8 r9, r12, r9
sel r6, r12, r11 ; hev mask: r6
;vp8_filter() function
; load soure data to r6, r11, r12, lr
ldrh r9, [src, #-2]
ldrh r10, [src], pstep
pkhbt r12, r7, r8, lsl #16
ldrh r7, [src, #-2]
ldrh r8, [src], pstep
pkhbt r11, r9, r10, lsl #16
ldrh r9, [src, #-2]
ldrh r10, [src], pstep
; Transpose needs 8 regs(r6 - r12, and lr). Save r6 and lr first
str r6, [sp]
str lr, [sp, #4]
pkhbt r6, r7, r8, lsl #16
pkhbt lr, r9, r10, lsl #16
;transpose r12, r11, r6, lr to r7, r8, r9, r10
TRANSPOSE_MATRIX r12, r11, r6, lr, r7, r8, r9, r10
;load back hev_mask r6 and filter_mask lr
ldr r12, c0x80808080
ldr r6, [sp]
ldr lr, [sp, #4]
eor r7, r7, r12 ; p1 offset to convert to a signed value
eor r8, r8, r12 ; p0 offset to convert to a signed value
eor r9, r9, r12 ; q0 offset to convert to a signed value
eor r10, r10, r12 ; q1 offset to convert to a signed value
str r9, [sp] ; store qs0 temporarily
str r8, [sp, #4] ; store ps0 temporarily
str r10, [sp, #8] ; store qs1 temporarily
str r7, [sp, #12] ; store ps1 temporarily
qsub8 r7, r7, r10 ; vp8_signed_char_clamp(ps1-qs1)
qsub8 r8, r9, r8 ; vp8_signed_char_clamp(vp8_filter + 3 * ( qs0 - ps0))
and r7, r7, r6 ; vp8_filter (r7) &= hev (r7 : filter)
qadd8 r7, r7, r8
ldr r9, c0x03030303 ; r9 = 3 --modified for vp8
qadd8 r7, r7, r8
ldr r10, c0x04040404
qadd8 r7, r7, r8
;mvn r11, #0 ; r11 == -1
and r7, r7, lr ; vp8_filter &= mask
;modify code for vp8 -- Filter1 = vp8_filter (r7)
qadd8 r8 , r7 , r9 ; Filter2 (r8) = vp8_signed_char_clamp(vp8_filter+3)
qadd8 r7 , r7 , r10 ; vp8_filter = vp8_signed_char_clamp(vp8_filter+4)
mov r9, #0
shadd8 r8 , r8 , r9 ; Filter2 >>= 3
shadd8 r7 , r7 , r9 ; vp8_filter >>= 3
shadd8 r8 , r8 , r9
shadd8 r7 , r7 , r9
shadd8 lr , r8 , r9 ; lr: filter2
shadd8 r7 , r7 , r9 ; r7: filter
;usub8 lr, r8, r10 ; s = (s==4)*-1
;sel lr, r11, r9
;usub8 r8, r10, r8
;sel r8, r11, r9
;and r8, r8, lr ; -1 for each element that equals 4 -- r8: s
;calculate output
;qadd8 lr, r8, r7 ; u = vp8_signed_char_clamp(s + vp8_filter)
ldr r8, [sp] ; load qs0
ldr r9, [sp, #4] ; load ps0
ldr r10, c0x01010101
qsub8 r8, r8, r7 ; u = vp8_signed_char_clamp(qs0 - vp8_filter)
qadd8 r9, r9, lr ; u = vp8_signed_char_clamp(ps0 + Filter2)
;end of modification for vp8
eor r8, r8, r12
eor r9, r9, r12
mov lr, #0
sadd8 r7, r7, r10
shadd8 r7, r7, lr
ldr r10, [sp, #8] ; load qs1
ldr r11, [sp, #12] ; load ps1
bic r7, r7, r6 ; r7: vp8_filter
qsub8 r10 , r10, r7 ; u = vp8_signed_char_clamp(qs1 - vp8_filter)
qadd8 r11, r11, r7 ; u = vp8_signed_char_clamp(ps1 + vp8_filter)
eor r10, r10, r12
eor r11, r11, r12
sub src, src, pstep, lsl #2
;we can use TRANSPOSE_MATRIX macro to transpose output - input: q1, q0, p0, p1
;output is b0, b1, b2, b3
;b0: 03 02 01 00
;b1: 13 12 11 10
;b2: 23 22 21 20
;b3: 33 32 31 30
; p1 p0 q0 q1
; (a3 a2 a1 a0)
TRANSPOSE_MATRIX r11, r9, r8, r10, r6, r7, r12, lr
strh r6, [src, #-2] ; store the result
mov r6, r6, lsr #16
strh r6, [src], pstep
strh r7, [src, #-2]
mov r7, r7, lsr #16
strh r7, [src], pstep
strh r12, [src, #-2]
mov r12, r12, lsr #16
strh r12, [src], pstep
strh lr, [src, #-2]
mov lr, lr, lsr #16
strh lr, [src], pstep
|vskip_filter|
sub src, src, #4
subs count, count, #1
ldrne r6, [src], pstep ; load source data
ldrne r7, [src], pstep
ldrne r8, [src], pstep
ldrne lr, [src], pstep
bne Vnext8
add sp, sp, #16
ldmia sp!, {r4 - r11, pc}
ENDP ; |vp8_loop_filter_vertical_edge_armv6|
;-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
|vp8_mbloop_filter_vertical_edge_armv6| PROC
;-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
stmdb sp!, {r4 - r11, lr}
sub src, src, #4 ; move src pointer down by 4
ldr count, [sp, #40] ; count for 8-in-parallel
ldr r12, [sp, #36] ; load thresh address
pld [src, #23] ; preload for next block
sub sp, sp, #16 ; create temp buffer
ldr r6, [src], pstep ; load source data
ldrb r4, [r2] ; blimit
pld [src, #23]
ldr r7, [src], pstep
ldrb r2, [r3] ; limit
pld [src, #23]
ldr r8, [src], pstep
orr r4, r4, r4, lsl #8
ldrb r3, [r12] ; thresh
orr r2, r2, r2, lsl #8
pld [src, #23]
ldr lr, [src], pstep
mov count, count, lsl #1 ; 4-in-parallel
orr r4, r4, r4, lsl #16
orr r3, r3, r3, lsl #8
orr r2, r2, r2, lsl #16
orr r3, r3, r3, lsl #16
|MBVnext8|
; vp8_filter_mask() function
; calculate breakout conditions
; transpose the source data for 4-in-parallel operation
TRANSPOSE_MATRIX r6, r7, r8, lr, r9, r10, r11, r12
uqsub8 r7, r9, r10 ; p3 - p2
uqsub8 r8, r10, r9 ; p2 - p3
uqsub8 r9, r10, r11 ; p2 - p1
uqsub8 r10, r11, r10 ; p1 - p2
orr r7, r7, r8 ; abs (p3-p2)
orr r10, r9, r10 ; abs (p2-p1)
uqsub8 lr, r7, r2 ; compare to limit. lr: vp8_filter_mask
uqsub8 r10, r10, r2 ; compare to limit
sub src, src, pstep, lsl #2 ; move src pointer down by 4 lines
orr lr, lr, r10
uqsub8 r6, r11, r12 ; p1 - p0
uqsub8 r7, r12, r11 ; p0 - p1
add src, src, #4 ; move src pointer up by 4
orr r6, r6, r7 ; abs (p1-p0)
str r11, [sp, #12] ; save p1
uqsub8 r10, r6, r2 ; compare to limit
uqsub8 r11, r6, r3 ; compare to thresh
orr lr, lr, r10
; transpose uses 8 regs(r6 - r12 and lr). Need to save reg value now
; transpose the source data for 4-in-parallel operation
ldr r6, [src], pstep ; load source data
str r11, [sp] ; push r11 to stack
ldr r7, [src], pstep
str r12, [sp, #4] ; save current reg before load q0 - q3 data
ldr r8, [src], pstep
str lr, [sp, #8]
ldr lr, [src], pstep
TRANSPOSE_MATRIX r6, r7, r8, lr, r9, r10, r11, r12
ldr lr, [sp, #8] ; load back (f)limit accumulator
uqsub8 r6, r12, r11 ; q3 - q2
uqsub8 r7, r11, r12 ; q2 - q3
uqsub8 r12, r11, r10 ; q2 - q1
uqsub8 r11, r10, r11 ; q1 - q2
orr r6, r6, r7 ; abs (q3-q2)
orr r7, r12, r11 ; abs (q2-q1)
uqsub8 r6, r6, r2 ; compare to limit
uqsub8 r7, r7, r2 ; compare to limit
ldr r11, [sp, #4] ; load back p0
ldr r12, [sp, #12] ; load back p1
orr lr, lr, r6
orr lr, lr, r7
uqsub8 r6, r11, r9 ; p0 - q0
uqsub8 r7, r9, r11 ; q0 - p0
uqsub8 r8, r12, r10 ; p1 - q1
uqsub8 r11, r10, r12 ; q1 - p1
orr r6, r6, r7 ; abs (p0-q0)
ldr r7, c0x7F7F7F7F
orr r8, r8, r11 ; abs (p1-q1)
uqadd8 r6, r6, r6 ; abs (p0-q0) * 2
and r8, r7, r8, lsr #1 ; abs (p1-q1) / 2
uqsub8 r11, r10, r9 ; q1 - q0
uqadd8 r6, r8, r6 ; abs (p0-q0)*2 + abs (p1-q1)/2
uqsub8 r12, r9, r10 ; q0 - q1
uqsub8 r6, r6, r4 ; compare to flimit
orr r9, r11, r12 ; abs (q1-q0)
uqsub8 r8, r9, r2 ; compare to limit
uqsub8 r10, r9, r3 ; compare to thresh
orr lr, lr, r6
orr lr, lr, r8
mvn r11, #0 ; r11 == -1
mov r12, #0
usub8 lr, r12, lr
ldr r9, [sp] ; load the compared result
sel lr, r11, r12 ; filter mask: lr
cmp lr, #0
beq mbvskip_filter ; skip filtering
;vp8_hevmask() function
;calculate high edge variance
sub src, src, pstep, lsl #2 ; move src pointer down by 4 lines
orr r9, r9, r10
ldrh r7, [src, #-2]
ldrh r8, [src], pstep
usub8 r9, r12, r9
sel r6, r12, r11 ; hev mask: r6
; vp8_mbfilter() function
; p2, q2 are only needed at the end. Don't need to load them in now.
; Transpose needs 8 regs(r6 - r12, and lr). Save r6 and lr first
; load soure data to r6, r11, r12, lr
ldrh r9, [src, #-2]
ldrh r10, [src], pstep
pkhbt r12, r7, r8, lsl #16
ldrh r7, [src, #-2]
ldrh r8, [src], pstep
pkhbt r11, r9, r10, lsl #16
ldrh r9, [src, #-2]
ldrh r10, [src], pstep
str r6, [sp] ; save r6
str lr, [sp, #4] ; save lr
pkhbt r6, r7, r8, lsl #16
pkhbt lr, r9, r10, lsl #16
;transpose r12, r11, r6, lr to p1, p0, q0, q1
TRANSPOSE_MATRIX r12, r11, r6, lr, r7, r8, r9, r10
;load back hev_mask r6 and filter_mask lr
ldr r12, c0x80808080
ldr r6, [sp]
ldr lr, [sp, #4]
eor r7, r7, r12 ; ps1
eor r8, r8, r12 ; ps0
eor r9, r9, r12 ; qs0
eor r10, r10, r12 ; qs1
qsub8 r12, r9, r8 ; vp8_signed_char_clamp(vp8_filter + 3 * ( qs0 - ps0))
str r7, [sp, #12] ; store ps1 temporarily
qsub8 r7, r7, r10 ; vp8_signed_char_clamp(ps1-qs1)
str r10, [sp, #8] ; store qs1 temporarily
qadd8 r7, r7, r12
str r9, [sp] ; store qs0 temporarily
qadd8 r7, r7, r12
str r8, [sp, #4] ; store ps0 temporarily
qadd8 r7, r7, r12 ; vp8_filter: r7
ldr r10, c0x03030303 ; r10 = 3 --modified for vp8
ldr r9, c0x04040404
;mvn r11, #0 ; r11 == -1
and r7, r7, lr ; vp8_filter &= mask (lr is free)
mov r12, r7 ; Filter2: r12
and r12, r12, r6 ; Filter2 &= hev
;modify code for vp8
;save bottom 3 bits so that we round one side +4 and the other +3
qadd8 r8 , r12 , r9 ; Filter1 (r8) = vp8_signed_char_clamp(Filter2+4)
qadd8 r12 , r12 , r10 ; Filter2 (r12) = vp8_signed_char_clamp(Filter2+3)
mov r10, #0
shadd8 r8 , r8 , r10 ; Filter1 >>= 3
shadd8 r12 , r12 , r10 ; Filter2 >>= 3
shadd8 r8 , r8 , r10
shadd8 r12 , r12 , r10
shadd8 r8 , r8 , r10 ; r8: Filter1
shadd8 r12 , r12 , r10 ; r12: Filter2
ldr r9, [sp] ; load qs0
ldr r11, [sp, #4] ; load ps0
qsub8 r9 , r9, r8 ; qs0 = vp8_signed_char_clamp(qs0 - Filter1)
qadd8 r11, r11, r12 ; ps0 = vp8_signed_char_clamp(ps0 + Filter2)
;save bottom 3 bits so that we round one side +4 and the other +3
;and r8, r12, r10 ; s = Filter2 & 7 (s: r8)
;qadd8 r12 , r12 , r9 ; Filter2 = vp8_signed_char_clamp(Filter2+4)
;mov r10, #0
;shadd8 r12 , r12 , r10 ; Filter2 >>= 3
;usub8 lr, r8, r9 ; s = (s==4)*-1
;sel lr, r11, r10
;shadd8 r12 , r12 , r10
;usub8 r8, r9, r8
;sel r8, r11, r10
;ldr r9, [sp] ; load qs0
;ldr r11, [sp, #4] ; load ps0
;shadd8 r12 , r12 , r10
;and r8, r8, lr ; -1 for each element that equals 4
;qadd8 r10, r8, r12 ; u = vp8_signed_char_clamp(s + Filter2)
;qsub8 r9 , r9, r12 ; qs0 = vp8_signed_char_clamp(qs0 - Filter2)
;qadd8 r11, r11, r10 ; ps0 = vp8_signed_char_clamp(ps0 + u)
;end of modification for vp8
bic r12, r7, r6 ;vp8_filter &= ~hev ( r6 is free)
;mov r12, r7
;roughly 3/7th difference across boundary
mov lr, #0x1b ; 27
mov r7, #0x3f ; 63
sxtb16 r6, r12
sxtb16 r10, r12, ror #8
smlabb r8, r6, lr, r7
smlatb r6, r6, lr, r7
smlabb r7, r10, lr, r7
smultb r10, r10, lr
ssat r8, #8, r8, asr #7
ssat r6, #8, r6, asr #7
add r10, r10, #63
ssat r7, #8, r7, asr #7
ssat r10, #8, r10, asr #7
ldr lr, c0x80808080
pkhbt r6, r8, r6, lsl #16
pkhbt r10, r7, r10, lsl #16
uxtb16 r6, r6
uxtb16 r10, r10
sub src, src, pstep, lsl #2 ; move src pointer down by 4 lines
orr r10, r6, r10, lsl #8 ; u = vp8_signed_char_clamp((63 + Filter2 * 27)>>7)
qsub8 r8, r9, r10 ; s = vp8_signed_char_clamp(qs0 - u)
qadd8 r10, r11, r10 ; s = vp8_signed_char_clamp(ps0 + u)
eor r8, r8, lr ; *oq0 = s^0x80
eor r10, r10, lr ; *op0 = s^0x80
strb r10, [src, #-1] ; store op0 result
strb r8, [src], pstep ; store oq0 result
mov r10, r10, lsr #8
mov r8, r8, lsr #8
strb r10, [src, #-1]
strb r8, [src], pstep
mov r10, r10, lsr #8
mov r8, r8, lsr #8
strb r10, [src, #-1]
strb r8, [src], pstep
mov r10, r10, lsr #8
mov r8, r8, lsr #8
strb r10, [src, #-1]
strb r8, [src], pstep
;roughly 2/7th difference across boundary
mov lr, #0x12 ; 18
mov r7, #0x3f ; 63
sxtb16 r6, r12
sxtb16 r10, r12, ror #8
smlabb r8, r6, lr, r7
smlatb r6, r6, lr, r7
smlabb r9, r10, lr, r7
smlatb r10, r10, lr, r7
ssat r8, #8, r8, asr #7
ssat r6, #8, r6, asr #7
ssat r9, #8, r9, asr #7
ssat r10, #8, r10, asr #7
sub src, src, pstep, lsl #2 ; move src pointer down by 4 lines
pkhbt r6, r8, r6, lsl #16
pkhbt r10, r9, r10, lsl #16
ldr r9, [sp, #8] ; load qs1
ldr r11, [sp, #12] ; load ps1
ldr lr, c0x80808080
uxtb16 r6, r6
uxtb16 r10, r10
add src, src, #2
orr r10, r6, r10, lsl #8 ; u = vp8_signed_char_clamp((63 + Filter2 * 18)>>7)
qsub8 r8, r9, r10 ; s = vp8_signed_char_clamp(qs1 - u)
qadd8 r10, r11, r10 ; s = vp8_signed_char_clamp(ps1 + u)
eor r8, r8, lr ; *oq1 = s^0x80
eor r10, r10, lr ; *op1 = s^0x80
ldrb r11, [src, #-5] ; load p2 for 1/7th difference across boundary
strb r10, [src, #-4] ; store op1
strb r8, [src, #-1] ; store oq1
ldrb r9, [src], pstep ; load q2 for 1/7th difference across boundary
mov r10, r10, lsr #8
mov r8, r8, lsr #8
ldrb r6, [src, #-5]
strb r10, [src, #-4]
strb r8, [src, #-1]
ldrb r7, [src], pstep
mov r10, r10, lsr #8
mov r8, r8, lsr #8
orr r11, r11, r6, lsl #8
orr r9, r9, r7, lsl #8
ldrb r6, [src, #-5]
strb r10, [src, #-4]
strb r8, [src, #-1]
ldrb r7, [src], pstep
mov r10, r10, lsr #8
mov r8, r8, lsr #8
orr r11, r11, r6, lsl #16
orr r9, r9, r7, lsl #16
ldrb r6, [src, #-5]
strb r10, [src, #-4]
strb r8, [src, #-1]
ldrb r7, [src], pstep
orr r11, r11, r6, lsl #24
orr r9, r9, r7, lsl #24
;roughly 1/7th difference across boundary
eor r9, r9, lr
eor r11, r11, lr
mov lr, #0x9 ; 9
mov r7, #0x3f ; 63
sxtb16 r6, r12
sxtb16 r10, r12, ror #8
smlabb r8, r6, lr, r7
smlatb r6, r6, lr, r7
smlabb r12, r10, lr, r7
smlatb r10, r10, lr, r7
ssat r8, #8, r8, asr #7
ssat r6, #8, r6, asr #7
ssat r12, #8, r12, asr #7
ssat r10, #8, r10, asr #7
sub src, src, pstep, lsl #2
pkhbt r6, r8, r6, lsl #16
pkhbt r10, r12, r10, lsl #16
uxtb16 r6, r6
uxtb16 r10, r10
ldr lr, c0x80808080
orr r10, r6, r10, lsl #8 ; u = vp8_signed_char_clamp((63 + Filter2 * 9)>>7)
qadd8 r8, r11, r10 ; s = vp8_signed_char_clamp(ps2 + u)
qsub8 r10, r9, r10 ; s = vp8_signed_char_clamp(qs2 - u)
eor r8, r8, lr ; *op2 = s^0x80
eor r10, r10, lr ; *oq2 = s^0x80
strb r8, [src, #-5] ; store *op2
strb r10, [src], pstep ; store *oq2
mov r8, r8, lsr #8
mov r10, r10, lsr #8
strb r8, [src, #-5]
strb r10, [src], pstep
mov r8, r8, lsr #8
mov r10, r10, lsr #8
strb r8, [src, #-5]
strb r10, [src], pstep
mov r8, r8, lsr #8
mov r10, r10, lsr #8
strb r8, [src, #-5]
strb r10, [src], pstep
;adjust src pointer for next loop
sub src, src, #2
|mbvskip_filter|
sub src, src, #4
subs count, count, #1
pld [src, #23] ; preload for next block
ldrne r6, [src], pstep ; load source data
pld [src, #23]
ldrne r7, [src], pstep
pld [src, #23]
ldrne r8, [src], pstep
pld [src, #23]
ldrne lr, [src], pstep
bne MBVnext8
add sp, sp, #16
ldmia sp!, {r4 - r11, pc}
ENDP ; |vp8_mbloop_filter_vertical_edge_armv6|
; Constant Pool
c0x80808080 DCD 0x80808080
c0x03030303 DCD 0x03030303
c0x04040404 DCD 0x04040404
c0x01010101 DCD 0x01010101
c0x7F7F7F7F DCD 0x7F7F7F7F
END
|
